This is a bit complicated. Even with the cyclic transition from child to child, there is a short period after the child loses focus and before the other child gains focus, where not a single child from a complex widget has focus. Although this condition is extremely short-lived, it is a fact.
The only parameter that I see is to decide how long the period between the child widgets loses focus and the absence of the child widget that you think is long enough to state that the child widget is actually no longer in focus. This will be a mechanism similar to what is used to highlight two mouse clicks with one click.
, , . (100 ?) , . , , . , , , .